JP6061496B2 - Pattern measuring apparatus, pattern measuring method, and pattern measuring program - Google Patents

Description

  The present invention relates to a measuring apparatus and a measuring method for measuring a dimension of a fine pattern created on a sample using an electron microscope.

  In the manufacturing process of a semiconductor device in which a number of fine circuit patterns are formed, it is measured whether the pattern is processed as designed or whether there is a variation in pattern dimensions. For this pattern measurement, a scanning electron microscope (SEM) suitable for measuring the dimensions of a fine pattern is widely used. In SEM, a scanning image (SEM image) is obtained by irradiating a measurement sample with an electron beam while scanning and detecting signal electrons (secondary electrons and reflected electrons) obtained secondarily from the measurement sample. Since the amount of signal electrons generated secondarily depends on the pattern shape and material, the measurement target pattern can be measured using the contrast of the SEM image. For example, Patent Documents 1 to 3 disclose a measuring apparatus and a measuring method using an electron microscope.

JP 2004-140280 A Japanese Patent Laid-Open No. 7-27549 JP 2009-110969 A

  A semiconductor device is generally composed of a plurality of layers, but both upper layer and lower layer patterns may be simultaneously captured in an SEM image during manufacture. For example, when a part of the upper layer is removed by etching or the like and the lower layer is exposed. In this case, since the electron beam is irradiated to the exposed lower layer in addition to the upper layer, the SEM image includes both the upper layer pattern and the exposed lower layer pattern. When the upper layer and lower layer patterns are mixed in the SEM image, there is an advantage that the upper layer and the lower layer can be measured simultaneously, and the positional deviation between the upper layer and the lower layer can be measured.

  However, when the upper layer and lower layer patterns are mixed in the SEM image, accurate measurement may be hindered. When the upper layer pattern is a measurement target, if the contour of the lower layer pattern is close to the contour of the upper layer pattern, the pattern contour to be measured may be erroneous. For example, a case will be described where the lower layer wiring appears as a line pattern in the hole opening by measuring the hole pattern. FIG. 1 and FIG. 2 show examples in which a hole pattern is present in the upper layer and a line pattern is present in the lower layer. A part of the line pattern 11 is exposed in the hole opening 10. In FIG. 1, a part of the outline 11 a of the line pattern can be seen inside the hole opening outline 10. In this case, when the hole opening contour 10 is measured, the contour 11a of the line pattern may be erroneously set as a measurement target. Further, the appearance of the line pattern 11 may change. In FIG. 2, both the outlines 11a and 11b of the line pattern are exposed in the hole opening. For this reason, if the contour extraction is performed on the assumption that only one side of the lower layer wiring contour is exposed as shown in FIG. 1, the contour of the measurement target may be erroneous. Such a difference in appearance is caused by process variations such as dimensional variations in holes and lower layer wiring and misalignment between upper and lower layers, and therefore cannot always be predicted from pattern design.

  The necessity to eliminate the erroneous determination of the contour has increased with the recent miniaturization of semiconductor devices. Conventionally, the design dimension can be provided with a margin, but since the miniaturization advances and the design dimension cannot be provided with a margin, the process variation is easily realized. For example, in the example of FIG. 1, by making the wiring in the lower layer of the hole sufficiently thicker than the hole, it is possible to suppress the outline of the lower wiring from reaching the inside of the hole due to process variations. However, in recent semiconductor devices, the wiring width and the hole diameter may be substantially equal, and the wiring outline easily appears or does not appear inside the hole due to process variations.

  In addition, a measurement method has been proposed in which the upper and lower patterns are simultaneously measured due to the manifestation of process variations. For example, it is necessary to measure misalignment between the upper layer and the lower layer and to measure an upper layer pattern on a specific pattern in the lower layer. As an example of the latter, there is an example in which the gate electrode on the channel region most important for the characteristics of the transistor is measured in the sample after processing the gate electrode of the transistor.

  On the other hand, with the miniaturization of the semiconductor device, it has become difficult to perform the above-mentioned contour discrimination by SEM. This is because it is necessary to discriminate a contour approaching several nm or less, which is equivalent to the resolution of SEM. In general, in SEM, the contour of a pattern is determined from the change in luminance of signal electrons. However, when the pattern contour is close to several nanometers or less, it is difficult to determine which change is due to the overlap of luminance changes. Even when the upper layer and the lower layer are measured simultaneously, the measurement may be difficult due to the miniaturization of the semiconductor device. This is because the interval between the patterns is narrowed, and the contour extraction in the upper and lower layers is likely to be hindered by the contours of the patterns in different layers.

  Patent Document 1 discloses a method of distinguishing and correctly measuring patterns of each layer when the SEM image includes patterns of a plurality of layers. The outline of the pattern is extracted from the SEM image and grouped to correctly determine the outline and perform measurement. However, as described above, it is difficult to extract a plurality of contours to be determined when the contours are close to each other. Therefore, this technique cannot distinguish between adjacent contours.

  Patent Document 2 discloses a method of measuring upper and lower layer patterns by changing electron beam irradiation conditions. According to Patent Document 2, a structure embedded in a sample can be measured by irradiating an electron beam having energy that can reach a portion that is not exposed to the electron beam. Furthermore, multiple SEM images are acquired using multiple irradiation conditions with different depths that can reach the sample and multiple irradiation conditions with different incident angles on the sample, and the three-dimensional shape is evaluated using the multiple SEM images. can do. Further, a method of performing image calculation such as laminating images of respective depths or creating an image including a luminance difference between SEM images has been proposed.

  However, these techniques cannot distinguish and measure the close contours of the upper layer and the lower layer. In the method of evaluating a three-dimensional shape by stacking SEM images having different depths, patterns with sufficiently different depths from the surface can be distinguished. However, the pattern existing on substantially the same plane is the same as the measurement by the conventional SEM image. In general, when the upper layer and the lower layer are measured simultaneously, attention is paid to the surface where the upper layer and the lower layer are in contact. This is because the difference in wiring and device structure between the upper layer and the lower layer affects the performance and yield of the semiconductor device. For this reason, the contours of the upper layer and the lower layer pattern to be measured are often on substantially the same plane. For example, in FIG. 1, the hole bottom contour 10 and the contours 11 a and 11 b of the lower layer wiring exposed at the hole opening are substantially in the same plane of the interface between the upper layer and the lower layer. In addition, it is not possible to distinguish and measure adjacent contours simply by creating an image for measurement by performing image computation such as a difference image created by a luminance difference between images. The adjacent contours are unclear because of overlapping luminance changes. For this reason, in the image obtained by the image calculation, the calculation error becomes large, and it is impossible to distinguish whether it is a contour to be measured truly or a contour that is accidentally generated by calculating an indefinite luminance change. In addition, since the imaging conditions are different, a plurality of SEM images cannot be acquired simultaneously. When the imaging is not performed at the same time, the imaging position may be shifted even if the sample position is not intentionally changed. This is because the sample moves due to vibration or thermal vibration, the charged state of the sample changes, and the irradiation position of the electron beam shifts. Therefore, with these techniques, it is impossible to distinguish and measure a contour close to several nanometers.

  Patent Document 3 discloses a technique for obtaining SEM images for measurement by acquiring SEM images with respect to a single measurement object under a plurality of imaging conditions and combining these images. Patent Document 3 discloses a technique for acquiring a plurality of images with different focal positions with respect to a hole pattern and synthesizing images that are in focus at both the top and bottom of the hole. With these techniques, the possibility of erroneously measuring the adjacent contours as in the past cannot be excluded for the adjacent contours existing on substantially the same plane. For example, an image of the top of the hole is not necessary for measuring the bottom of the hole, and it is difficult to determine the contour of the bottom of the hole with the combined image of the top and the bottom.

  An object of the present invention is to provide a pattern measurement device, a pattern measurement method, and a pattern measurement program capable of correctly selecting and measuring a measurement target pattern contour even if the pattern contours are close to each other in a sample including a plurality of patterns. It is.

  In order to achieve the above object, the present invention has the structure described in the claims.

  To give an example of the pattern measurement apparatus of the present invention, a pattern is captured in a detection image by scanning a sample with charged particles and detecting secondary charged particles or reflected charged particles generated from the sample. An image acquisition unit that acquires a plurality of detection images having different imaging conditions at substantially the same location of a sample, and an outline extraction unit that extracts a plurality of pattern contours from the plurality of detection images A contour reconstruction unit that reconstructs a measurement target contour by combining the plurality of pattern contours, and a contour measurement unit that performs measurement using the reconstructed measurement target contour.

  In addition, if an example of the pattern measurement method of the present invention is given, a sample is scanned with charged particles, a secondary charged particle or a reflected charged particle generated from the sample is detected, a detection image is created, and the detection image is captured. A pattern measurement method for measuring a pattern, a step of acquiring a plurality of detection images having different imaging conditions at substantially the same location of a sample, a step of extracting a plurality of pattern contours from the plurality of detection images, The method includes a step of reconfiguring a measurement target contour by combining a plurality of pattern contours, and a step of measuring using the measurement target contour.

  An example of the pattern measurement program of the present invention is a program for causing a computer to perform measurement of a pattern imaged on a detection image obtained by a charged particle beam apparatus, in substantially the same place of a sample. A step of reading a plurality of detection images having different imaging conditions, a step of extracting a plurality of pattern contours from the plurality of detection images, a step of reconstructing a measurement target contour by combining the plurality of pattern contours, and the measurement Measuring using the target contour.

  According to the present invention, in a sample including a plurality of patterns, the measurement target pattern contour can be correctly selected and measured even when the pattern contours are close to each other.

The figure which shows an example of the SEM image of the sample which has a hole pattern in an upper layer, and a line pattern in a lower layer. The figure which shows another example of the SEM image of the sample which has a hole pattern in an upper layer, and a line pattern in a lower layer. The flowchart explaining the content of the pattern measurement of Example 1 of this invention. The block block diagram of the pattern measuring device of Example 1 of this invention. The figure explaining the sample made into the measuring object in Example 1 of this invention. In Example 1 of this invention, the figure which shows an example of the SEM image of the sample made into the measuring object. In Example 1 of this invention, the figure which shows another example of the SEM image of the sample made into the measuring object. FIG. 3 is a diagram for explaining an example of a method for storing a plurality of SEM images in one image in Embodiment 1 of the present invention. The flowchart explaining the pattern shape measurement (S8) of FIG. 3 in detail. FIG. 3 is a diagram for explaining contour extraction of a measurement target pattern from an SEM image in Embodiment 1 of the present invention. The brightness | luminance profile of the SEM image explaining the outline extraction of FIG. FIG. 3 is a diagram for explaining contour extraction of a measurement target pattern from an SEM image in Embodiment 1 of the present invention. The brightness | luminance profile of the SEM image explaining the outline extraction of FIG. FIG. 3 is a diagram showing two types of pattern contours extracted from two types of SEM images in Example 1 of the present invention. The figure which shows the measurement object pattern reconfigure | reconstructed from two types of pattern outlines extracted from two types of SEM images in Example 1 of this invention. The figure explaining the example of the measuring object outline determination method performed in step S23 of FIG. FIG. 3 is a diagram illustrating an example of a GUI screen used for pattern measurement in the first embodiment of the present invention. The flowchart explaining the content of the pattern measurement of Example 2 of this invention. FIG. 10 is a diagram illustrating an example of a GUI screen used for pattern measurement in the second embodiment of the present invention. The figure which shows an example of the SEM image of the sample which has a line pattern in an upper layer and a lower layer. The flowchart which deform | transformed a part of pattern measurement of Example 1 of this invention. The flowchart explaining the content of the pattern measurement of Example 3 of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. As embodiments for carrying out the present invention, two modes will be described. In the first embodiment, a new SEM capable of implementing the present invention is used. In the second mode, an image is acquired with a conventional SEM, and then measurement is performed with a pattern measurement program capable of implementing the present invention.

  In the first embodiment, the imaging conditions of the SEM image, the contour extraction method, and the pattern measurement method can be collectively managed by the SEM. Therefore, there are advantages that the entire processing time can be shortened and that handling such as parameter setting is easy. On the other hand, in the second embodiment, since the conventional SEM can be used, the effects of the present invention can be obtained at a lower cost than in the first embodiment.

  FIG. 4 is a block diagram of the pattern measuring apparatus according to the first embodiment. The pattern measurement apparatus 100 used in this embodiment includes an image acquisition unit 101, a calculation unit 102, a data storage device 103, and a display device 104. The image acquisition unit 101 only needs to have a configuration capable of acquiring an SEM image. For example, the image acquisition unit 101 has the same configuration as a CD-SEM (Critical Dimension Scanning Electron Microscope). Device control parameters and data processing parameters for imaging and measuring a measurement sample are stored in the data storage device 103 as a wafer measurement recipe. The wafer measurement recipe includes the position of the measurement target pattern, reference images and data necessary for pattern matching so that the measurement target falls within the imaging range, setting conditions such as an electron beam necessary for imaging, pattern measurement method and measurement Includes the necessary data processing conditions. The calculation unit 102 is configured by a computer, controls the operation of each unit, and performs measurement on the captured SEM image. The calculation unit 102 includes a misalignment detection unit 1021, a contour extraction unit 1022, a contour reconstruction unit 1023, and a contour measurement unit 1024. When the measurement is started, the calculation unit 102 reads the wafer measurement recipe from the data storage device 103, and adjusts the sample position necessary for acquiring the image of the measurement target pattern and the control parameters necessary for imaging the image. To give. The image acquisition unit 101 executes image acquisition and returns the obtained image to the calculation unit 102. The computing unit 102 performs image processing and pattern shape measurement specified in the wafer measurement recipe on the received image. For the measurement of the pattern shape, a misalignment detection unit 1021, a contour extraction unit 1022, a contour reconstruction unit 1023, and a contour measurement unit 1024 are used. Further, the image and pattern measurement results are stored in the data storage device 103. In addition, the calculation unit 102 displays a series of processing steps on the display device 104 as appropriate.

  The measurement target sample of Example 1 will be described with reference to FIG. The sample to be measured in this example is in a state where a multilayer wiring of a semiconductor device is being created. FIG. 5 schematically shows an example of an SEM image obtained from the sample to be measured. From FIG. 5, three types of patterns can be observed in the SEM image of the sample to be measured. That is, the trench pattern 106, the hole pattern 107 processed inside the trench pattern 106, and the lower layer wiring pattern 108 inside the hole pattern 107. Reference numeral 105 denotes a mask film. Explaining the laminated structure, the trench pattern 106 is formed on the outermost surface, and the hole pattern 107 is processed from the bottom of the trench pattern 106. Under the hole pattern 107, a lower wiring pattern 108 is formed. The trench pattern 106 and the hole pattern 107 are hollow patterns that are processed only in shape, but the lower wiring pattern 108 is filled with metal.

  5 is a part of the metal wiring 109 exposed by processing the trench pattern 106 and the hole pattern 107. The lower wiring pattern 108 shown in FIG. In FIG. 5, the right outline of the lower layer wiring 108 can be seen inside the hole pattern 107. This is because the positions of the lower layer wiring 109 and the hole pattern 107 are shifted. This shift varies depending on the manufacturing process. For example, if the center of the metal wiring 109 and the center of the hole pattern 107 are formed to coincide with each other, the entire bottom surface of the hole pattern 107 becomes a metal wiring, and the outline of the metal wiring cannot be seen inside the hole.

  The measurement target in this example was the diameter of the hole pattern 107 shown in FIG. 5 and the area of the lower wiring pattern 108 exposed inside the hole. The present invention is not limited to the case where the pattern is exposed on the surface. One or both of the plurality of patterns may be buried in the sample.

  Next, the pattern measurement procedure in the present embodiment will be described. In this embodiment, two SEM images are acquired using two types of imaging conditions (first imaging condition and second imaging condition) for one measurement location. If the sample has multiple measurement locations, two types of procedures are possible. The first procedure is a method of acquiring images of the first imaging condition and the second imaging condition continuously after adjusting the sample position so that an SEM image of the measurement location can be acquired. The second procedure is a method of acquiring SEM images of all measurement locations of the sample under the first imaging condition, and then acquiring SEM images of all measurement locations of the sample again under the second imaging condition. In the first procedure, the sample needs to be aligned only once, but the apparatus condition may not be stable because the imaging conditions are changed in a short time. In the second procedure, since the sample is aligned under each of the first imaging condition and the second imaging condition, the measurement time becomes long. However, since the imaging conditions need only be switched once, it is easy to stabilize the apparatus state of the SEM. In this example, after acquiring SEM images under the first imaging condition at all measurement locations of the latter sample, SEM images of all measurement locations of the sample were acquired under the second imaging condition.

  Based on FIG. 3, the outline of the flow of the pattern measurement of Example 1 is demonstrated. Details of each step will be described later. First, a sample is set in the apparatus (step S1). Next, SEM images of all sample measurement locations are captured under the first imaging condition. In order to adjust the sample position and the irradiation range position of the electron beam so that the measurement target enters the imaging region, the first sample position adjustment (step S2) is performed. Next, a first SEM image is acquired under the first imaging condition and stored in the data storage device 103 (step S3). Steps S2 and S3 are repeated at each measurement location of the sample. Next, SEM images of all measurement locations of the sample are captured under the second imaging condition. Second sample position adjustment (step S4) and acquisition of a second SEM image under the second imaging condition (step S5) are performed. Next, the shift amount of the imaging position is detected from the first SEM image and the second SEM image captured at the same measurement location (step S6). Next, the first SEM image and the second SEM image are associated with each measurement location of the sample (step S7). Here, to associate is to store the first and second SEM images at the same imaging location and the imaging position deviation amounts of both SEM images in the data storage device 103 as data that can be easily extracted. Next, the pattern shape is measured using the first SEM image and the second SEM image (step S8). Steps S4 to S8 are repeated for n measurement locations on the sample. Note that the present invention does not depend on the order of measurement procedures. For example, a parallel operation can be performed in order to speed up the measurement. For example, step S4 and step S5 of the steps in FIG. 3 can be repeatedly performed by the image acquisition unit 101, and the calculation unit 102 can repeatedly perform steps S6 to S7 in parallel.

  The parameters necessary for proceeding from step S1 to step S8 in FIG. 3 are created in advance as a wafer measurement recipe. For this reason, after the sample is placed in the apparatus in step 1, steps S <b> 1 to S <b> 8 can be automatically advanced by causing the calculation unit 102 to execute the wafer measurement recipe.

  Next, the first imaging condition and the second imaging condition will be described. Here, the imaging conditions include, for example, the incident energy and current value of the electron beam incident on the sample, the angle at which the electron beam is incident on the sample, the tip diameter of the electron beam, the beam opening angle, the scanning speed of the electron beam, These are the scanning order, scanning direction, extraction voltage applied in the direction of extracting signal electrons from the sample surface, and the like. If the incident energy or incident angle of the electron beam is changed, the depth at which the electron beam reaches the sample changes. For example, the greater the incident energy or the perpendicular to the sample, the deeper the electron beam can enter the sample. Therefore, it is possible to measure a buried pattern without being exposed on the surface. On the other hand, under the reverse condition, it is easy to measure the pattern on the sample surface. It is also possible to enhance the secondary electron image that easily reflects the shape of the sample by reducing the incident energy of the electron beam, and to emphasize the reflected electron image that easily reflects the material of the sample by increasing the incident energy. It is. Further, if the opening angle of the electron beam is reduced to make the beam narrow, the electron beam can reach the bottom of the pattern through a finer pattern and a gap between the patterns. On the other hand, if the beam has a large opening angle and a thick beam, the pattern on the sample surface can be easily measured without entering the gap between the fine patterns. Further, the charged state of the sample can be changed by the electron beam current, the scanning speed and scanning order of the electron beam, and the extraction voltage. This is because these parameters can adjust the amount of electrons incident on the sample and the amount of electrons extracted from the sample.

  When the charged state is changed between the sample surface and the pattern bottom, the potential difference between the sample surface and the pattern bottom changes, and secondary electrons generated at the pattern bottom can be suppressed or emphasized. As a means for changing the charged state of the sample surface, pre-dose in which an electron beam is irradiated in advance before an image for measurement is acquired is also effective. The charged state of the sample depends not only on the electron beam irradiation method but also on the material of the sample, the laminated structure, and the processed pattern structure. Therefore, the above parameters are adjusted so that a desired SEM image is obtained according to the measurement target.

  The first imaging condition is an imaging condition that can easily measure the shape of the hole pattern to be measured and can also image the trench pattern on the sample surface. The second imaging condition is an imaging condition in which the metal wiring pattern below the hole pattern can be easily measured and the trench pattern on the sample surface can be imaged. In this embodiment, the incident energy of the electron beam is 500 V, the current is 2 pA, and the scanning speed is increased under the first imaging condition. Further, flat scan was used as the electron beam scanning procedure. Here, the flat scan is a method of scanning in an arbitrary direction discretely for each coordinate point. Under this imaging condition, it is possible to enhance the secondary electron image and capture the first SEM image in which the hole shape can be easily measured.

  FIG. 6 shows a schematic diagram of an SEM image obtained by applying the first imaging condition to the same pattern as in FIG. The hole pattern 107 is emphasized, and the contour of the lower layer wiring 108 is weaker than the contour of the hole pattern 107. Under the second imaging condition, the incident energy of the electron beam was increased and the current was increased. 1600V and 16pA, respectively. Also, the scanning speed was slow. The scanning method was interlaced scanning. Here, the interlaced scanning is a scanning method in which one row is skipped in the vertical direction of the imaging range, the horizontal scanning is repeated, and the horizontal scanning of the next skipped column is repeated. As a result, the second SEM image can be a SEM image in which the reflected electron image is more emphasized than the first SEM image and the metal wiring can be easily measured. FIG. 7 shows a schematic diagram of an SEM image obtained under the second imaging condition for the same pattern as in FIG. Contrary to FIG. 6, the outline of the hole pattern 107 is weak and the outline of the lower layer wiring 108 is emphasized. In both the first and second SEM images, the imaging conditions were set so that the trench pattern of the first sample layer could be imaged. This is because it is used for detection of the imaging position shift amount of the first and second SEM images, which will be described in detail later.

  FIG. 7 is also an example in which the imaging position is deviated, and the position of the hole pattern 107 is shifted to the right side from FIG. In the present invention, the imaging conditions can be arbitrarily set according to the material and structure to be measured.

  Next, detection of the imaging position deviation amount in step S6 will be described. In this example, sample position adjustment was performed before the first and second SEM image acquisition. Therefore, the imaging position of the SEM image is shifted according to the sample position adjustment accuracy. Further, although different from the present embodiment, even when the SEM image is acquired continuously under the first and second imaging conditions after the sample position adjustment, the sample is caused by vibration or thermal vibration between the two imaging operations. There is a possibility that the irradiation range of the electron beam may be shifted due to a shift or a change in the charged state of the sample. Therefore, when obtaining one measurement value using both of these two types of SEM images, it is necessary to correct the amount of imaging position deviation between the SEM images.

  In this example, the trench pattern on the sample surface that is not the measurement target was imaged in both the first SEM image and the second SEM image. Using this commonly captured trench pattern, the amount of imaging position deviation between both images was detected. First, the outline of the trench pattern is extracted from the first and second SEM images. Next, the correlation value between images was obtained while shifting the positions of the extracted contours, and the shift amount with the largest correlation value was taken as the imaging position shift amount of both images. In the direction crossing the trench, the outline of a plurality of trenches appears, so that the positional deviation can be detected clearly. Even in a direction parallel to the trench, the roughness of the trench is rough, so that the amount of imaging position deviation can be detected by pattern matching. Since the roughness of the contour is inevitably generated due to the material of the trench and the manufacturing process, it is possible to detect the amount of positional deviation stably by this method.

  As a method for detecting the amount of displacement, an optimum method can be selected in accordance with the pattern imaging state. For example, it is possible to use a method of performing Fourier transform on the luminance change of an image such as a phase only correlation method. Further, not only the positional deviation due to the parallel movement but also the deviation in the rotation direction can be corrected. As a method for detecting a shift in the rotation direction, for example, there is a method using a rotation invariant phase-only correlation. Further, it is also possible to detect the amount of displacement after performing noise removal and contour enhancement / extraction on the image. The present invention does not depend on the alignment method. In the present embodiment, the imaging position deviation amount is detected using a pattern different from the measurement target pattern. This is because the visibility of the measurement target pattern is greatly different between the first and second SEM images, and the error increases when the amount of imaging position deviation is detected using the detection target pattern. However, if it is suitable for detecting the amount of displacement, the imaging position may be adjusted including the hole pattern and lower layer wiring pattern that are the measurement target.

  Next, step S7 for associating the first SEM image and the second SEM image for each measurement location will be described. In this step, data including the data names of the two SEM images, the measurement location coordinates, the displacement amount, and the like may be created and recorded in the data storage device 103. Also, two images can be integrated. Integration refers to making two pieces of image data into one piece of image data. At this time, it is desirable that each of the two pieces of image data can be easily reproduced, and that the image is a corrected image position deviation. For example, it is integrated into an image format corresponding to RGB three colors. The SEM image is generally a monochromatic image in order to use the detected signal electron intensity as luminance information. Therefore, if the luminance information of the SEM image is stored in each of the three colors RGB, it is possible to store up to three images. Alternatively, it is integrated into a format that supports frames and multi-pages. For example, in the moving image format, a plurality of images can be stored as frames. Even in an image format, a multi-page TIFF or animation GIF format can store a plurality of images.

  In this embodiment, the images are integrated into an image corresponding to RGB three colors. As shown in FIG. 8, the luminance of the first SEM image is stored as the luminance of R of the three RGB colors, and the luminance of the second SEM image is stored as the luminance of G. B was left empty. Further, in the second SEM image stored in the luminance of G, in order to eliminate the imaging position deviation from the first SEM image, based on the pattern imaging position deviation amount detected from the first measurement image in step S6. The pixel position was shifted. The image size was made larger than the captured image so as to complement the amount of positional deviation. In this integrated image, if only R is displayed, the first SEM image can be displayed, and if only G is displayed, the second EM image in which the imaging position is combined with the first SEM image can be displayed.

  Next, the pattern shape measurement S8 will be described. In the case of normal CD-SEM measurement, generally, a pattern contour to be measured is extracted and the distance between the contours is measured. In the present invention, a plurality of SEM images with different imaging conditions are acquired for one measurement location. Therefore, it is possible to extract the contour of the measurement target pattern from each image and perform measurement on the contour, or to extract the contour from a calculation result image between SEM images such as a difference image and a sum image and perform measurement. The latter is suitable when the contour of the measurement target pattern can be clarified by performing computation between images. However, since an error occurs in the calculation of the image, the former is suitable when measurement accuracy is important. In this way, an optimum contour extraction method including combining a plurality of images can be selected according to the measurement target and the image quality of the SEM image. Further, the contour to be measured can be selected from a plurality of contours extracted from a plurality of images. For example, the amount of deviation between the contour extracted from the first SEM image and the contour extracted from the second SEM image can be measured.

  Next, details of the procedure of the pattern shape measurement S8 will be described with reference to FIG. First, contours are extracted from the first and second SEM images (steps S21 and S22). In order to extract the contour of the pattern from the image, for example, the brightness of the SEM image, the derivative of the brightness, the change point of the second derivative of the brightness, the maximum value, the minimum value, or any position between the maximum value and the minimum value Is detected as a contour point. A contour is obtained by fitting a straight line, a curve, or a combination thereof to a plurality of contour points. In order to facilitate contour extraction, image processing such as noise removal and contour enhancement can be performed on the SEM image in advance. In addition, an optimum contour extraction method can be used for each imaging condition. Note that the present invention does not depend on a method for extracting a contour from an SEM image.

  Next, the contour extraction shown in FIGS. 10 and 11 will be described. In this embodiment, the outline of the hole pattern is extracted from the first SEM image. First, the contour extraction target range 110 is provided.

  As shown in FIG. 10, the center of this range is set to overlap the approximate center of the hole. The approximate center of the hole is obtained from pattern matching with a reference image registered in the recipe. Next, the contour is determined by paying attention to the luminance change in the contour extraction target range 110.

  FIG. 11 shows a luminance profile at the location indicated by the arrow in FIG. In the first SEM image, the imaging conditions are optimized so that the hole shape can be easily measured, and the luminance is greatly reduced inside the hole as shown in FIG. There is also a change in luminance due to the metal wiring inside the hole, but it is smaller than the luminance change at the hole outline. Therefore, the hole contour can be determined without being obstructed by the contour of the metal wiring. After smoothing the luminance profile of FIG. 11, the luminance in the contour extraction target range 110 is searched from the approximate center of the hole toward the outside of the hole, and the maximum value and the minimum value are determined. In FIG. 11, when the luminance is searched from the center to the left side, a position A ′ where the luminance is minimum and a position A where the luminance is maximum are found. Next, the luminance at A is 100%, the luminance at A ′ is 0%, and the luminance position A ″ corresponding to 40% is set as the contour point of the hole. Similarly, the luminance was searched to the right from the center of the hole, B at the maximum luminance position and B ′ at the minimum position were found, and B ″ at the 40% position was determined as the contour point of the hole. Similar processing is performed in a direction to rotate 180 ° with respect to the center of the hole, and the contour point of the whole hole is determined. Next, an approximate curve or straight line is obtained from the set of contour points to obtain a contour. The contour extracted from the first SEM image is an ellipse that approximates the set of extracted contour points.

  Next, contour extraction from the second SEM image will be described. The contour of the lower metal wiring is extracted from the second SEM image.

  First, as shown in FIG. 12, a contour extraction target range 110 is provided near the hole. The position of the contour extraction target range 110 is determined by pattern matching with the reference image in the same manner as the first SEM image. Since the second SEM image is different from the first SEM image in the hole outline and the brightness of the lower layer wiring, the reference image for determining the approximate center of the hole is an image different from the reference image used in the first SEM image. It is desirable that

  FIG. 13 shows a luminance profile at the location indicated by the arrow in FIG. In the second SEM image, the imaging conditions were optimized so that the lower layer metal wiring can be easily measured. In FIG. 13, the luminance of the metal wiring is higher than that in FIG. As for the contour of the metal wiring, the brightness in the contour extraction target range 110 is searched from the approximate center of the hole to the outside of the hole, and the contour point is set at the position where the brightness is minimized. In FIG. 13, the position A ″ where the luminance is searched for from the approximate center of the hole to the left and the luminance is minimized and the position B ″ where the luminance is searched for and minimized from the right are used as the contour points of the metal wiring. A similar process was performed in a direction rotated by 180 ° with respect to the center of the hole, and the contour point of the entire metal wiring was determined. Next, a curve or straight line approximated to the set of extracted contour points was obtained, and a contour composed of an ellipse and a straight line was obtained as shown by the hatched portion contour in FIG.

  Next, the procedure (step S23) for determining the measurement target contour from the contours extracted from the first and second SEM images shown in FIG. 9 will be described. In this step, a contour obtained by combining pattern contours extracted from a plurality of images can be reconstructed and used as a measurement target. FIG. 14 shows the contour extracted in this embodiment. In FIG. 14, the contour 112 extracted from the first SEM image and the contour 111 extracted from the second SEM image are displayed in an overlapping manner. In this embodiment, the diameter of the hole pattern 107 and the area of the lower layer wiring pattern 108 were measured. For the former, the contour 112 extracted from the first SEM image was used as the measurement target contour. For the latter, as shown by the hatched area in FIG. 15, the region included in both the contour 112 extracted from the first SEM image and the contour 111 extracted from the second SEM image is used as the measurement target pattern. The contour was taken as the measurement target contour.

  The reason why the contour extracted from the first and second SEM images is used for the measurement of the lower wiring pattern 108 will be described. The second SEM image is imaged under an imaging condition optimized to capture the contour of the metal wiring pattern exposed in the hole. However, when the outline of the metal wiring is hidden without being exposed in the hole, the outline becomes relatively unclear.

  This can also be seen from the fact that the change in luminance at the boundary A ″ between the metal wiring and the hole is smaller than the contour B ″ of the metal wiring exposed in the hole in FIG. In addition, the second SEM image capturing condition is that the incident energy of the electron beam on the sample is relatively high. Therefore, when the insulating film is thin in the vicinity of the hole boundary, there is a possibility that signal electrons caused by the metal wiring that is not exposed through the electron beam are captured. For this reason, in the second SEM image, the boundary between the metal wiring and the hole may appear outside the actual side.

  14 and 15, the contour of the wiring pattern 108 extracted from the second SEM image does not coincide with the contour 112 extracted from the first SEM image, and there is a portion located outside. For this reason, it is considered that the boundary between the hole and the metal wiring can be more accurately extracted from the first SEM image. Therefore, a region included in the contour 112 extracted from the first SEM image and included in the contour 111 extracted from the second SEM image is set as a measurement target pattern. The contour of the measurement target pattern was reconstructed as the contour of the measurement target.

  The present invention does not depend on the contour reconstruction method. A reconstruction method can be selected according to the measurement target and the image to be captured. As another reconstruction method, for example, as shown in FIG. 16, the outline of the region (1-2) included in the first SEM image and not included in the second SEM image, or the first SEM image This is the outline of the area to be included and the area (1 + 2) to be included in the second SEM image.

  In step S23, when a plurality of contours are extracted from the SEM image, it is possible to select a contour to be reflected in the measurement target contour. For example, in the second SEM image shown in FIG. 7, the case where both the contour of the hole pattern 107 and the contour of the lower layer wiring 108 are detected using a contour extraction method different from the present embodiment will be described. In step S23, conditions are determined so that the contour of the second SEM image is selected with priority given to the contour having a straight line. Therefore, it is possible to select the outline of the hatched area with many straight lines among the outlines detected from FIG. Thereafter, as described above, the contour of the measurement target pattern is reconstructed.

  Similarly, when a plurality of contours are extracted from the SEM image, it is possible to select a contour to be reflected in the measurement target contour by comparing with a contour extracted from another SEM image.

  For example, in the second SEM image shown in FIG. 7, the case where both the contour of the hole pattern 107 and the contour of the lower layer wiring 108 are detected using a contour extraction method different from the present embodiment will be described. In step S23, a selection condition is determined so that a contour different from the contour extracted from the first SEM image is preferentially selected. Compared with the contour of the hole pattern 107 obtained from the first SEM image of FIG. 6, the contour of the lower layer wiring 108 which is a different contour is selected. As a method of selecting a different contour, for example, there is a method of selecting a contour having a large deviation amount of the center of gravity of the contour.

  Further, FIG. 21 shows a procedure for performing contour extraction again using another contour extraction method based on a result of comparing the contour extracted from the first SEM image and the contour extracted from the second SEM image. In FIG. 21, step S25 is provided in which the contour extracted from the first SEM image and the contour extracted from the second SEM image are compared and determined. As a result of the comparison of both contours, for example, if it is determined that different contours are obtained, the process proceeds to step S23. Conversely, if it is determined that the same contour has been selected, the process returns to step S21 and the contour extraction is performed again.

  As described above, the method of combining the contour information extracted from each of the first SEM image and the second SEM image and then reflecting the information to the contour of each image is to capture a plurality of SEM images with different imaging conditions. This is effective when the contour cannot be clearly distinguished only by the contour extraction method suitable for each image.

  Further, as described with reference to FIG. 8, in the present embodiment, the amount of imaging position deviation from the first SEM image is corrected in the second SEM image. In the case where the correction of the imaging position deviation amount is not reflected in the image and only the imaging position deviation amount data is left, the imaging position deviation amount of the image detected in step S8 is reflected in the contour position in this step.

  As described above, a procedure of capturing a plurality of images with different imaging conditions, extracting a contour by an extraction method suitable for each image, aligning image capturing positions, and determining a measurement target contour from a plurality of contours By passing through, it is possible to correctly select the measurement target contour without confusion between adjacent contours.

  Next, the measurement target contour is measured (step S24). In this step, the feature amount of the pattern shape such as the length of the contour, the contour interval, the inclination, the angle, and the area is quantified from the target contour. In this embodiment, the diameter of the hole pattern 107 and the area of the lower wiring pattern 108 are to be measured. For the diameter of the hole pattern 107, an ellipse that approximates the contour 112 extracted from the first SEM image was obtained, and the average diameter was used as the measurement value. For the area of the lower layer wiring pattern 108, the area included in the outline of the hatched area in FIG. The present invention does not depend on the method for quantifying the target contour.

  Next, a GUI necessary for performing the procedure described based on FIG. 9 will be described. In order to perform the procedure described with reference to FIG. 9, it is possible to set conditions for each of a contour extraction method for each of a plurality of SEM images, a method for determining a measurement target contour from a plurality of contours, and a method for performing measurement using a measurement target contour. A GUI is required.

  FIG. 17 is an example of a GUI for creating a measurement recipe necessary for performing the procedure of FIG. Here, the measurement recipe is a parameter set necessary for performing the procedure of FIG. The data display unit 201 includes a pattern recipe list display unit 202, a contour extraction method setting unit 203, a contour reconstruction method setting unit 204, and a pattern measurement method setting unit 205. The relationship between each setting unit and the steps in FIG. 9 is as follows. The contour extraction method setting unit 203 sets necessary parameters in steps S21 and S22 in FIG. 9, the contour reconstruction method setting unit 204 in step S23 in FIG. 9, and the pattern measurement method setting unit 205 sets necessary parameters in step S24 in FIG.

  Next, an outline of a measurement recipe creation procedure will be described. In this embodiment, a plurality of measurements can be registered for one measurement place. A measurement number is assigned to each measurement and displayed on the measurement number display unit 211. Next, in each measurement, necessary parameters are set in the contour extraction method setting unit 203, the contour reconstruction method setting unit 204, and the pattern measurement method setting unit 205, and the registration button 206 is pressed. When the registration button 206 is pressed, the above parameters for each measurement are collectively stored as a pattern recipe. At this time, the calculation unit 102 temporarily stores it. The stored pattern recipe is displayed on the pattern recipe list display unit 202. The information on the pattern recipe displayed on the display on the pattern recipe list display unit 202 includes, for example, a measurement number, an SEM image number to be used, a contour extraction method, a contour reconstruction method, and a pattern measurement method. After all necessary measurements are registered, the save button 208 is pressed. The save button 208 collectively saves one or a plurality of pattern recipes displayed on the pattern recipe list display unit 202 as a measurement recipe. At this time, the name of the measurement recipe is given and the measurement recipe is stored in the data storage device 103. The name of the measurement recipe is displayed on the measurement recipe display unit 210. Further, by pressing a read button 207, a measurement recipe stored in the data storage device 103 can be called to the arithmetic unit 102. Furthermore, when the execution button 209 is pressed, a window for selecting a measurement target image appears. When a measurement target image is selected, a measurement recipe called by the calculation unit 102 can be executed on the image. By including the measurement recipe in the wafer measurement recipe, the same measurement is performed at all measurement locations of the sample according to the procedure of FIG.

  The details of the GUI in FIG. 17 will be described below in accordance with the procedure shown in FIG. In this example, two types of measurements were performed: the diameter of the hole pattern 107 and the area of the lower wiring pattern 108. First, the case where the latter lower wiring pattern 108 is a measurement pattern will be described. First, an example of a measurement target SEM image or design data is read as a reference image. Recipe creation proceeds while actually executing processing on this reference image.

  The contour extraction method setting unit 203 sets whether or not to use a plurality of SEM images at the same measurement location for contour extraction, and if so, the contour extraction conditions. In this embodiment, the contour of the lower wiring pattern 108 is extracted from two sheets of the first SEM image and the second SEM image. As shown in FIG. 15, both the contour 112 extracted from the first SEM image and the contour 111 extracted from the second SEM image were used. Therefore, each contour extraction method is set by the contour extraction method setting unit 203. First, the contour extraction SEM image selection unit 2031 is set to select and use the first SEM image (SEM image number 1). At this time, the first SEM image of the reference image is displayed on the data display unit 201. Further, the contour extraction parameter setting unit 2032 sets a contour extraction method. The contour extraction method set here is the method described with reference to FIG. When set, the contour extracted from the first SEM image of the reference image by the set extraction method is superimposed on the first SEM image of the reference image and displayed on the data display unit 201. Since the present invention does not depend on the details of the contour extraction parameter setting unit 2032, the description of the contour extraction parameter setting unit 2032 is omitted. Next, the contour extraction SEM image selection unit 2031 selects and sets the second SEM image (SEM image number 2) to be used. At this time, the second SEM image of the reference image is displayed on the data display unit 201. Further, the contour extraction parameter setting unit 2032 sets a contour extraction method. The contour extraction method set here is the method described with reference to FIG. When the contour extraction method is set, the contour extracted from the second SEM image of the reference image by the set extraction method is superimposed on the second SEM image of the reference image and displayed on the data display unit 201. Although there is only one contour extraction parameter setting unit 2032 in the contour extraction method setting unit 203, different parameters can be set for each image.

  Next, the contour reconstruction method setting unit 204 will be described. Here, the settings necessary for the operation of step S23 in FIG. 9 are performed. In the measurement of the lower layer wiring pattern 108, as described above, a region included in both the contour 112 extracted from the first SEM image and the contour 111 extracted from the second SEM image was used as a measurement target pattern. In the contour reconstruction method setting unit 204, the reconstruction target SEM image selection unit 2041 selects the first SEM image (SEM image number 1) and the second SEM image (SEM image number 2). Further, the reconstruction method selection unit 2042 selects “AND”. “AND” is registered as a method for obtaining a region included in the contour extracted from the image selected by the synthesis target SEM image selection unit 2041 and reconstructing the contour. The present invention does not depend on the contour reconstruction method. A necessary contour reconstruction method can be created and selected from the selection unit 2042. When the setting items are set, the reconfigured contour is displayed on the data display unit 201.

  Next, the pattern measurement method setting unit 205 sets the measurement method described in step 24. For the contour of the hatched area in FIG. 15 reconstructed in step S23, setting is made so that the inclusion area becomes a measured value. Since the present invention does not depend on the details of the pattern measurement method setting unit 205, the description of the pattern measurement method setting unit 205 is omitted. When this setting is completed, the measured value is displayed on the data display unit 201.

  Next, a setting method using the GUI of FIG. 17 will be described for the measurement of the diameter of the hole pattern 107, which is another measurement of the present embodiment. Here, only the difference from the above-described measurement of the area of the lower layer wiring pattern 108 will be briefly described. For the diameter of the hole pattern 107, the contour is extracted only from the first SEM image, and no reconstruction is performed. Accordingly, it is set that the first SEM image (SEM image number 1) is selected and used by the contour extraction SEM image selection unit 2031 and the contour extraction method described with reference to FIG. 11 is set by the contour extraction parameter setting unit 2032. To do. Next, the SEM image selection unit 2031 selects the second SEM image (SEM image number 2) and sets it to not used. In the contour reconstruction method setting unit 204, the selection unit 2041 selects the first SEM image (SEM image number 1), and the reconstruction method selection unit 2042 selects “no reconstruction”. “No reconstruction” is registered as a method in which the contour extracted from the SEM image is directly used as the measurement target contour. The pattern measurement method setting unit 205 obtains an ellipse that approximates the contour, and sets the average diameter as a measurement value.

  In the present invention, it does not depend on the structure of the sample to be measured. In this example, a sample was used in which the trench pattern 106, the hole pattern 107, and the lower layer wiring pattern 108 described in FIG. 5 can be observed. The present invention is also effective for samples having different structures.

  FIG. 20 shows an example in which both the upper layer and the lower layer are line patterns. For example, in the case of a sample after processing the gate electrode of a transistor, the gate electrode of the transistor is the upper layer line pattern 14, the element isolation region of the transistor is the lower line pattern 12, and the active region constituting the transistor channel is between the lower line patterns. The image of FIG. 20 is obtained as the space 13. In the image of FIG. 20, if the lower line pattern 12 is clearly visible, the measurement of the upper line pattern 14 is hindered. This is because the outline of the line pattern 14 and the outline of the lower layer line pattern 12 are confused at the place where the edge of the line pattern 12 and the edge of the line pattern 14 intersect. Therefore, in the same manner as in the present embodiment, by using the imaging conditions that can emphasize each of the upper layer line pattern 14 and the lower layer line pattern 12, the alignment of the images, and the contour extraction and measurement method optimized for each image. The outline of the line pattern 14 and the outline of the lower layer line pattern 12 can be measured without confusion.

  According to the present embodiment, it is possible to obtain an SEM image in which imaging conditions are optimized so that a pattern contour can be easily extracted for each of a plurality of adjacent patterns. Furthermore, an optimum contour extraction method can be adopted for each of a plurality of adjacent patterns. Thus, the contour of the target pattern can be extracted without being obstructed by other adjacent patterns. Furthermore, the pattern to which the contour belongs can be distinguished by tracing the extraction source SEM image. In addition, by performing SEM image imaging position alignment, it is possible to eliminate measurement errors due to imaging position shifts.

  Thus, in the step of selecting the contour of the measurement target, it is possible to prevent confusion between the contour of the measurement target pattern and the contour of another pattern that is close. In addition, the process of extracting the contour from the SEM image and the process of selecting the contour to be measured can be made independent, so even when measuring with contours belonging to multiple patterns, a contour extraction method suitable for each pattern is adopted. it can. Therefore, even if a plurality of patterns having greatly different visibility in the SEM image are targets, measurement can be performed correctly.

  In the second embodiment, an example in which an SEM image is acquired using a conventional SEM and the present invention is realized using a new pattern measurement program will be described. The pattern measurement program of the present embodiment is incorporated in a computer and causes the computer to operate as a pattern measurement device. By incorporating the pattern measurement program of this embodiment into a computer, for example, the calculation unit of FIG. 4 is configured. The measurement target sample used in this example is the same as that in Example 1. Moreover, the description which overlaps with Example 1 is omitted.

  First, a plurality of SEM images for the same imaging location are acquired using a conventional SEM. In the present example, two types of first SEM image and second SEM image were acquired. The imaging conditions for these images are the same as those for the first SEM image and the second SEM image in Example 1.

  Next, a procedure of processing performed by the pattern measurement program will be described with reference to FIG. FIG. 18 is a diagram showing an outline of a measurement procedure performed by the pattern measurement program used in this embodiment. First, the first and second SEM images are called up (step S31). Next, detection of an imaging position shift in the first and second SEM images (step S32), pattern contour extraction from the first SEM image (step S33), and pattern contour extraction from the second SEM image (step S33) Step S34) is performed. Next, a contour to be measured is determined from the contours obtained in steps S33 and S34 (step S35). In step S35, a contour to be measured is formed from the contour corrected for the imaging position deviation. Next, measurement is performed using the measurement target contour (step S36).

  The details of each step are the same as those described in the first embodiment of the present invention. Step S32 is the same as step S6 in FIG. 3, step S33 is the same as step S21 in FIG. 9, step S34 is the same as step S22 in FIG. 9, step S35 is the same as step S23 in FIG. 9, and step S36 is the same as step S24 in FIG.

  The present invention does not depend on the order of the steps in FIG. In FIG. 18, step S32, step S33, and step S34 are processed in parallel at the same time, but may be processed in succession.

  Next, a GUI required for the pattern measurement program used in this embodiment will be described. This program includes an imaging position shift detection method for a plurality of SEM images at the same measurement location, a contour extraction method for each of a plurality of SEM images, a method for determining a measurement target contour from a plurality of contours, and measurement using a measurement target contour. A GUI capable of setting conditions is required for each method to be performed. FIG. 19 shows an example of the GUI. This is an example in which an imaging position deviation detection method setting unit 212 is added to the GUI shown in FIG.

  The third embodiment is characterized in that, in addition to the procedure described in the first embodiment, the measurement result is determined and the imaging position alignment of a plurality of SEM images having different imaging conditions is performed again when it is determined as abnormal. This is effective when the imaging states of patterns commonly captured in SEM images with different imaging conditions are greatly different and it is difficult to align the imaging positions. Here, the case where imaging position alignment is difficult is, for example, a case where a plurality of shift amounts are detected based on the same determination criterion.

  FIG. 22 is a flowchart for explaining the features of this embodiment. In Example 1, the procedure shown in FIG. 9 was changed. Step S26 for determining the measurement value obtained in step S24 is provided to determine whether the measurement value is normal or abnormal. If it is determined to be normal, the flow ends. If it is determined that there is an abnormality, the process proceeds to step S27 in which the imaging position deviation amount is again performed. In step S27, the imaging position deviation amount is detected using a detection condition different from the imaging position deviation detection performed first. Next, the process returns to step S23. In step S23, the imaging position shift amount newly detected in step S27 is reflected in the position of the contour extracted from the SEM image. Furthermore, a new measurement target contour is determined using these contours. In step 24, measurement is performed using the newly determined measurement target contour. Next, the measured value is determined again in step S26. At this time, if the first measurement value and the measurement value obtained by re-detecting the imaging position deviation are the same, or if the number of times to repeat the imaging position deviation detection reaches a preset upper limit, an abnormality determination was made. The flow is finished with the measured value. In addition, the abnormality determination in step S26 is, for example, when the measured value exceeds a preset upper limit or lower limit, or when the same measurement is performed at a plurality of measurement locations, a difference from the average value of other measurement locations is preset. This is a determination that the upper limit or the lower limit is exceeded, or that the difference between the average value of past measurement values measured with the same shape sample and the same recipe exceeds the set upper limit or lower limit. Further, in this embodiment, not only the detection of the imaging position deviation amount but also the measurement contour determination method in step S26 can be redone from the determination of the measurement value.

  Furthermore, if the flow of FIG. 22 is used, for example, it is possible to automatically repeat the change of the imaging position detection method and the measurement target contour determination method so that the measurement value falls within a desired variation. With this method, it is possible to omit the time and effort to set the optimum imaging position detection method and measurement target contour determination method in advance, and to automatically set the conditions using the variation in the measurement values as an index.

DESCRIPTION OF SYMBOLS 10 Hole opening part 11 Line pattern 11a, 11b Line pattern outline 12 Lower layer line pattern 13 Lower layer line pattern space 14 Upper layer line pattern 100 Pattern measurement device 101 Image acquisition unit 102 Operation unit 103 Data storage device 104 Display device 105 Mask film 106 Trench pattern 107 Hole pattern 108 Lower layer wiring 109 Metal wiring 110 Outline extraction target range 111 Outline extracted from second SEM image 112 Outline extracted from first SEM image 201 Data display section 202 Pattern recipe list display section 203 Outline extraction method setting unit 204 Contour reconstruction method setting unit 205 Pattern measurement method setting unit 206 Registration button 207 Read button 208 Save button 209 Execution button 210 Measurement recipe display unit 211 Measurement number Display unit 212 Imaging position shift detection method setting unit 1021 Position shift detection unit 1022 Contour extraction unit 1023 Contour reconstruction unit 1024 Contour measurement unit 2031 SEM image selection unit 2032 Contour extraction parameter setting unit 2041 Compositing target SEM image selection unit 2042 Reconstruction method Select part

Claims (13)

A pattern measurement device that scans a sample with charged particles, creates a detection image by detecting secondary charged particles or reflected charged particles generated from the sample, and measures a pattern captured in the detection image,
An image acquisition unit that acquires a plurality of detection images having different imaging conditions at substantially the same location of the sample;
A contour extracting unit that extracts a plurality of pattern contours from the plurality of detected images;
A contour reconstruction unit that reconstructs a measurement target contour by combining the plurality of pattern contours;
A contour measurement unit that performs measurement using the reconfigured measurement target contour;
Equipped with a,
The imaging conditions include the incident energy and current value of the charged particles incident on the sample, the angle at which the charged particles are incident on the sample, the tip diameter and the opening angle of the charged particles, the scanning speed and scanning of the charged particles. A pattern measuring apparatus comprising an extraction voltage applied in an order or a scanning direction, or a direction in which the secondary charged particles or reflected charged particles are extracted from the sample surface . The pattern measurement apparatus according to claim 1,
A misregistration detection unit that detects an imaging misregistration using a pattern imaged in common to the plurality of detection images;
The contour reconstruction unit reconstructs a measurement target contour by reflecting the imaging position deviation. The pattern measurement apparatus according to claim 2,
The pattern displacement detection unit uses a pattern different from the measurement target pattern for detection of an imaging displacement. The pattern measurement apparatus according to claim 1,
A pattern measuring apparatus that stores a plurality of images in which imaging position shifts of the plurality of detected images are corrected as one image data. The pattern measurement apparatus according to claim 1,
The sample to be measured is composed of multiple layers,
A pattern measuring apparatus comprising a plurality of patterns to be measured and including at least one pattern belonging to different layers. The pattern measurement apparatus according to claim 1,
The pattern reconstruction apparatus, wherein the contour reconstruction unit includes a plurality of pattern contours in a pattern contour extracted from at least one image. The pattern measurement apparatus according to claim 1,
An interface for simultaneously displaying pattern contours extracted from a plurality of detection images acquired at substantially the same location of the sample;
The interface includes a contour reconstruction method setting unit. A pattern measurement method for scanning a sample with charged particles, creating a detection image by detecting secondary charged particles or reflected charged particles generated from the sample, and measuring a pattern captured in the detection image,
A step of acquiring a plurality of detected images where imaging conditions are different from each other in substantially the same location of the sample,
A step of extracting a plurality of pattern contours from said plurality of detected images,
A step of reconstructing the measurement object contour by combining the plurality of pattern edge,
And performing measurement by using a pre-Symbol instrumentation points contour,
Equipped with a,
The imaging conditions include the incident energy and current value of the charged particles incident on the sample, the angle at which the charged particles are incident on the sample, the tip diameter and the opening angle of the charged particles, the scanning speed and scanning of the charged particles. A pattern measurement method comprising: an extraction voltage applied in the order, the scanning direction, or the direction in which the secondary charged particles or reflected charged particles are extracted from the sample surface . The pattern measurement method according to claim 8,
A step of detecting an imaging position shift using a pattern imaged in common to the plurality of detection images,
Wherein the step of reconstructing the measurement object contour pattern measurement method characterized by reconstructing the measurement target contour to reflect the imaging position deviation. The pattern measurement method according to claim 9 , further comprising:
Determining whether the measurement value is normal or abnormal,
When it is determined that there is an abnormality, it includes a step of performing detection of an imaging position shift again,
Wherein the step of reconstructing the measurement object contour pattern measurement method comprising that you reconfigure the measurement target contour to reflect the re-detected imaging position deviation. The pattern measurement method according to claim 8 , further comprising:
Comparing and determining a plurality of extracted contours,
When it is determined that the same contour is selected, the pattern contour extraction is performed again in the step of extracting the pattern contour . A program for causing a computer to perform measurement of a pattern captured in a detection image obtained by a charged particle beam device,
Reading a plurality of detection images having different imaging conditions at substantially the same location of the sample;
Extracting a plurality of pattern contours from the plurality of detected images;
Reconstructing a measurement target contour by combining the plurality of pattern contours;
Performing measurement using the measurement target contour;
With
The imaging conditions include the incident energy and current value of the charged particles incident on the sample, the angle at which the charged particles are incident on the sample, the tip diameter and the opening angle of the charged particles, the scanning speed and scanning of the charged particles. A pattern measurement program comprising an extraction voltage applied in an order or a scanning direction, or a direction in which the secondary charged particles or reflected charged particles are extracted from the sample surface . In the pattern measurement program according to claim 12,
Detecting an imaging position shift using a pattern imaged in common to the plurality of detection images,
Wherein the step of reconfiguring the measurement object contour pattern measuring program reflects the previous SL IMAGING positional displacement, characterized in that reconstructing the measurement object contour.

Images